Download Macroeconomic Factors Thoroughbred Industry

Macroeconomic Factors
Thoroughbred Industry
Peter Karungu, Michael Reed and Douglas Tvedt*
Abstract
A capitalization approach is used to estimate econometrically the effects of exchange rate,
interest rate and tax law changes on thoroughbred yearling prices. The analysis found that exchange
rate and tax law changes have significantly influenced yearling prices since the early 1980s.
Another serious price-reducing event was the 1986 tax law change. Both of these factors have
counteracted the positive impact of increased purse rates on yearling prices.
Key Words: exchange rate, macroeconomics,
Agricultural economics research was tilted
towards macroeconomic analysis prior to the 1970s.
However, these macroeconomic analyses failed to
explain the decline in competitiveness of U.S.
In spite of
agriculture in the world market.
advanced technology and infrastructure and a wellestablished marketing system, U.S. competitiveness
in agriculture declined. This prompted Schuh’s
1974 article on the effects of the exchange rate on
U.S. agriculture.
He attributed lower U.S.
agricultural prices to an overvalued currency,
Schuh’s work ushered a research initiative which
concentrated on the impacts of the exchange rate on
U.S. agriculture (Kost; Chambers and Just; Johnson,
Grennes and Thursby). This literature has shown
that exchange rate changes have significantly
affected U.S. agriculture. Research efforts have
broadened in recent years to incorporate the effects
of macroeconomic variables -- interest rates,
inflation and changes in money supply -- on U.S.
agriculture.
This paper tests the hypothesis that
macroeconomic variables, variables which are not
tax laws, purse rates, thoroughbreds
specific to the thoroughbred industry, have
substantially affected the thoroughbred horse
industry in recent years. The thoroughbred industry
offers a unique opportunity to study the effects of
Thoroughbreds are
macroeconomic variables.
capital assets whose values stem fi-om expected
fiture net returns from racing and breeding, Many
thoroughbreds, especially yearlings 1, are sold at
auctions which bring together foreign and domestic
breeders. Thus, macroeconomic factors should play
a major role in determining the health of the
industry.
In the mid- 1980s, the thoroughbred
industry experienced an unprecedented economic
Foreign breeders invested heavily in
boom.
American-bred yearlings and this boosted prices to
historically high levels, especially at the major
thoroughbred sale sites where foreigners purchase
horses -- Keeneland (held in Lexington, Kentucky)
and Saratoga (held in Saratoga, New York) sales.
However, since 1985, the industry has undergone
significant economic decline. In 1984, the average
price of a selected yearling from the combined
*Karungu is a former research associate, Reed is a professor, and Tvedt is a research specialist in the department of
agricultural economics, University of Kentucky. Karungu is currently lecturer in the Department of Economics,
University of Witwatersrand University, Johannesburg, South Africa. The authors would like to thank Teresa Fitzgerald
for her assistance with data collection.
J. Agr. and Applied Econ. 25 (July, 1993): 165-173
Copyright 1993 Southern Agricultural Economics Association
166
Kurangu, Reed and Tvedt:
Keeneland and Saratoga summer sales was slightly
over $456,000 compared to an average of slightly
over $294,982 for 1990 (The Blood-Horse 1990).
The objective of this paper is to determine
the macroeconomic factors that have contributed to
this decline, Fundamental issues like the effects of
exchange rates, interest rates, and the 1986 Tax
Reform Act will be analyzed. Specifically,
objectives are:
1) To test hypotheses that the exchange
rate and the interest rate affect the price of
thoroughbred yearlings.
2) To provide quantitative measures of the
effects of these variables on year] ing prices
during the 1979-1990 period.
Literature Review
Chambers found a positive relationship
between the money supply and agricultural exports.
Orden expanded Chambers’ analysis to include
additional monetary variables (the money supply,
real interest rate, exchange rate and general price
level). He found that the exchange rate and real
interest rate explained the largest proportion of the
forecast-error variance for agricultural exports.
Bessler analyzed Brazilian data and concluded that
there was a one-way Granger-type causality from
money supply to agricultural prices.
Devados and Meyers tested the hypothesis
that farm output prices adjust faster than non-farm
output prices to monetary changes, Their results
confirmed that money supply increases (decreases)
benefit (harm) farmers because farm product prices
increase (decrease) relatively more than non-farm
product prices.
Findings from commodity-specific models
agree with findings from aggregated models that an
overvalued currency contributes to export declines.
Longmire and Morey, for example, concluded that
Macroeconomic
Factors
the appreciation of the exchange rate, approximately
20 percent during 1980-1982, caused the real value
of U.S. wheat, corn, and soybean exports to decline
by about $3 billion (or 16 percent). They argued
that this exchange rate effect accounted for
approximately three-fourths of the total decline in
the value of exports of these commodities during
this period.
The Industry
The equine industry in the U.S., with over
5.25 million horses and 32 states participating in
horse racing, is a multi-billion dollar industry
contributing over $15 billion annually to the gross
domestic product (Lawrence). Between 1975 and
1988, the industry contributed on average $600
million to the Federal government from its parimutuel wagering activities (Thoroughbred Times).
By 1988, 32 States had about 100 race tracks and a
total attendance of about 69 million. The industry,
through racing and breeding activities, contributes
significantly to the economies of many states.
The most visible and business-oriented
segment of the equine sector is the thoroughbred
industry, whose major role is to breed high-quality
horses for racing. Most offspring are auctioned as
yearlings at various sites throughout the country.
Sales at Keeneland and Saratoga are the most
publicized and offer the highest quality yearlings.
Between 1981 and 1990, 35 percent of the
total yearlings auctioned at Keeneland and Saratoga
This
were purchased by foreign investors.
accounted for 53 percent of the total gross
expenditure on yearling sales (The Blood-Horse).
Figure 1 illustrates the percentage of selected
yearlings that were purchased by foreign investors
and the percentage of the gross expenditures on
yearlings that accrued from the foreign market.
Obviously, foreign buyers tend to purchase higherpriced yearlings.
J. Agr. urrd Applied Ecrm., July, 1993
Figure
.
167
1.
Percent
70
60
50
40
30
20
10
0
1961
1982
1983
1964
1985
1966
1987
1988
1989
1990
Year
=
Source:
The
Blood
HorcIe,
Yearling Numbers
~~
Expenditures
1990
With few government restrictions on
thoroughbred purchasing and exporting (other than
quarantine), foreign buyers participate directly in the
auctions.
Thus, the exchange rate influences
thoroughbred auction prices directly. A stronger
dollar relative to other currmcies would raise
yearling prices for foreign buyers, while a weaker
dollar would generate demand for the high-quality
U.S. bred-yearlings.
The 1986 Tax Reform Act affected the
industry in three fundamental ways: I) The Passive
Activity Loss -- Prior to this Act, the thoroughbred
industry benefitted from capital infusions by people
who needed to shelter large sums of money, and
these investments in the industry were tax-exempt.
The Act put an end to these tax benefits. Newberry
argues that there is no change in the Internal
Revenue Code which has had such a profound
impact on the thoroughbred industry as the passive
loss provision. 2) The income tax from horses
classified as capital assets (mainly the broodmares
and stallions) rose from 20 percent to 28 percent,
and 3) the Act lowered the depreciation rate for the
industry.
to the buyer). The yearling price can be determined
by discounting the value of the expected future
earnings from racing and breeding.
Assume initially that all the expected
income generated by the yearling comes from U.S.
activities.
The price of the yearling can be
determined by discounting the nominal value of the
expected net future earnings from racing and
breeding in the U.S. Thus, applying an investment
model, the present value of a yearling may be
expressed as:
Where:
Conceptual Framework
V~ = price of the yearling
P, = expected net racing winnings in year t
T = tax rate
r
= nominal interest rate
R = number of years the horse can be expected to
race
B, = expected net income from breeding in year t
L = number of years the horse can be expected to
breed
Because yearlings are purchased for future
earnings, their price (value) can be modelled using
a capitalization approach. In this model, yearling
prices are demand-determined (i.e., from their value
This model assumes constant risk
premiums over time, no difference in tax rates
between ordinary income and capital gains, and no
salvage value for the thoroughbred. It abstracts
168
Karungu, Reed and Tvedt:
in price due to phenotypic factors
(genotypic and environmental factors) which
produce detectable variations in horses, While the
significance of the phenotypic factors cannot be
overlooked, they require critical evaluation of an
individual horse in regard to overall performance
and pedigree. These phenotypic factors are not
important for the present model due to the data
analyzed.
from variations
av
J=.
aE
aT
avd
<o,
t3r
avd
— >o,
aPO
avd
— >0
aBo
(2)
If the yearling is expected to earn all of its income
stream from overseas activities, the variables from
equation (1) would be:
v’.
R (1 -T*)P;
~
t.1
(I+r”)’
+ ~
‘;;:::
‘3)
~=
XavY__wv
E_lv_l
Y
vyaE-ff
(5)
If Wf is 1, Vy is V#E, and E~ is -1. That is, if all
horses were purchased for overseas activities (and
there
are
no substitutes
for
American
thoroughbreds), the elasticity of Vy with respect to
E would be -1.0. If Wf is O, Vy is Vd, and E~ is O.
That is, if all horses were purchased for domestic
activities, the elasticity of VY with respect to E
would be O. If arbitrage equates thoroughbred
prices between the U.S. and international markets,
..
If.
Ey
Where:
Vf = value of the yearling in foreign activities
P,* = expected net racing winnings in year t
T
= tax rate for foreign countries
l-” = nominal interest rate in foreign countries
R = number of years the horse can be expected to
race
B,* = expected net income from breeding in year t
L = number of years the horse can be expected to
breed
To transform equation (3) into dollars, one divides
the equation by E (denominated in foreign currency
per dollar).
Many horses are purchased with the
anticipation that they will be transported overseas
for all or part of their career, Thus, yearling prices
are a weighted average of equations (I) and (3):
vy=wdvd+w~f
wf7@
or
In equation (1) the following conditions
must hold:
<o;—
Fac[ors
V, is the value of the yearling from all activities in
all markets. Differentiating equation (4):
E
avd
Macroeconomic
fE
(4)
v
then
E~ . -W
f
,
Since many horses are purchased for U.S. activity,
the elasticity should fall between O and -1.
Model Specification and Data
The model is fitted with data from 1979 to
1990, The price data were collected from Keeneland
and Saratoga summer yearling sales (The BloodHorse). Data from each individual horse transaction
were placed into quartiles based on price, and the
median price from each quartile was used as the
dependent variable to eliminate phenotypic
variation, Thus, there are four observations per
year, The data on purse rate came from Keeneland
and were for the average per race in their spring
and fall meets.
Data for the exchange rate were obtained
from the Citibank Data Base. Data on the interest
rate were for intermediate loans from the Louisville
Farm Credit Services office. The exchange rate for
169
J. Agr. and Applied Ecrrn., July, 1993
each year was weighted by the proportion of sales
accounted for by each country. These countries, in
order of importance, are: England, Ireland, Canada,
France and Japan.
The sales to international
customers were obtained from a data base kept by
Keeneland (the purses are from Keeneland races),
The exchange rate is trade-weighted with 1982 as
the base3,
A dummy variable was used to identify the
years after the passage of the 1986 Tax Reform Act.
Another set of dummy variables identified the
quartile of the observation. Data on breeding prices
were not available.
;
P rote
The resulting regression model is a series of crosssectional observations over time. Each yearling is
sold only once, so there is only one time-series
observation for an individual cross-sectional unit,
Two estimates are obtained from the data - one for Keeneland and one for Saratoga. Separate
parameter estimates for the two sales allow an
analysis of differential variable effects that may
prevail for Keeneland and Saratoga. Because the
model is fitted in double logarithmic form, all
estimates are elasticities except for the coefficient
on dummy variables (e.g., T). The dummy variable
coefficients are elasticities if evaluated where the
dummy variable is unity.
E+
The results of both analyses are reported in
table 1. The Durbin-Watson statistic is 2.24 for the
top quartile, indicating that autocorrelation is not a
problem. All signs of the estimates conform to
theoretical expectations (except the coefficient for
the interest rate in the Keeneiand model) and most
slope estimates are significantly different from zero
at the 10 percent level. Both models explain over
96 percent of the variation in thoroughbred prices.
(6)
Where:
Py
a variable measuring tax laws,
= error term.
Results
Equation (4) cannot be transformed into a
functional
form which can be estimated
econometrically. Instead equations (1-4) were used
to identify the relevant right-hand-side variables.
For estimation purposes, the model was specified in
logarith@c form, which allows each right-hand-side
variable to have a constant elasticity. The model,
omitting dummy variables for the quartiles, is:
lnPY=bO +bllni+bzln
bd In P,,ate+ b6 T + u
T=
u
= yearling price,
= nominal interest rate,
= exchange rate,
=
purse rate for thoroughbred
races,
Table 1. Coefficient Estimates for the Yearling Price Model for Keeneland and Saratoga
Intercept
P,,lC
i
E
T
~2
D,
D,
D,
.1 ,39***
.,92***
-.50***
.97
,0s
.05
-.53***
.96
Keeneland
Coefficient
Standard
Error
-,776**
1.28**
-1,75***
0,031*
.0,29 ***
3.51
0.33
0.46
0.017”
0.09
-3.71
0.88*
-1,29**
-0.007
-0.17*
-1.43***
3.93
0.36
0.48
0.024
0.09
.05
Saratoga
Coefficient
Standard
Error
One asterisk impliessignificanceat the 10 percentlevel; two
at the 1 percent IeveL
implies significance at the 5
..97***
.05
!05
percent level; three implies significance
170
Keeneland
All coefficients for Keeneland data are
significantly different from zero at the 10 percent
level. The purse rate is an important determinant of
yearling prices during the observation period. In the
Keeneland model, a one percent increase in the
purse rate increases yearling prices by 1.28 percent,
though this coefficient estimate is not significantly
different than unity. This sensitivity to purse rate
changes could be capturing the expectation of future
purse rate increases over time. Not only will a one
percent increase in this year’s purse rate affect
yearling prices directly, but buyers might view this
as an indication that purse rates will increase in the
future.
The weighted exchange rate coefficient is
-1.75, which is outside the expected range for the
coefficient
(though the coefficient is not
significantly different from -1,0 at the 10 percent
level), The magnitude of this coefficient shows the
importance of foreigners in the pricing of yearling
thoroughbreds -- essentially price trends are set by
the foreign demand for yearlings. It is possible that
the market overreacts to exchange rate changes
because the demand by international buyers is so
important in the higher-priced yearlings, which set
the tone for the auction sales. A time-series
approach could detect such an overreaction.
Yearling prices at Keeneland are less
responsive to the interest rate than to other
variables, The coefficient for the interest rate is
actual Iy positive and significant y different from
zero at the 10 percent level, The interest rate
variable may be picking up variation associated with
an omitted variable which is positively linked to
prices (e.g., international yearling supplies and
inflation). Attempts to incorporate some of those
other variables (particularly income and the foal
crop size) into the regression model result in a
negative relationship between thoroughbred prices
and interest rates, but resulted in unexpected signs
on other coefficients (particularly the impact of
purse rates on thoroughbred prices). Obviously, the
independent variables are correlated in some of
these other models. Further, the intermediate term
loan from the Farm Credit System (FCS) is much
less volatile than many interest rates, such as
government bonds.
During 1990, the FCS
Karungn, Reed and Tvedt:
Macroeconomic
Factars
intermediate rate was still at 11.16 percent versus a
three-month treasury bill interest rate at 7.50
percent. Thus, the FCS interest rate may not be
picking up the true opportunity costs of funds. It is
difficult to determine which interest rate is most
relevant to the thoroughbred industry.
The dummy variable (which identifies
years when the 1986 Tax Reform Act was in force)
has a coefficient of -0.29, which means that the tax
law resulted in a 29 percent reduction in yearling
prices in 1986 (using 1986 values as a base).
Obviously, the passage of that Act was a crushing
blow to the industry. The elimination of tax
shelters and non-taxable investment as “passive
activities” have contributed substantially to the
decline in the industry.
Saratoga
All coefficients for the Saratoga model
have the expected signs, but all coefficients have
lower t-values than for Keeneland. The absolute
value of all coefficients are lower for Saratoga,
implying that yearling prices tend to be less volatile.
The quartile dummy variables explain more of the
variation for the Saratoga observations.
The effect of increases in the purse rate is
slightly smaller for the Saratoga model than the
Keeneiand model, but still increased purse rates
have a large impact on thoroughbred prices. A one
percent increase in the purse rate will increase
Saratoga yearling prices by 0.88 percent. This
coefficient estimate is significantly different from
zero at the 10 percent level, but not significantly
different from one.
The weighted exchange rate coefficient is
smaller in absolute value for Saratoga than for
Keeneland, -1,29 versus -1.75, but still outside the
expected range (though, again, the -1.29 coefficient
is not significantly different from - 1.0). Foreign
buyers are also quite important for the Saratoga
sales. The larger magnitude in the Keeneland
model is consistent with observations made by
prominent industry members that there are more
foreign buyers of yearlings at Keeneland than at
Saratoga. However, both models are extremely
sensitive to exchange rate changes.
171
J. Agr. und Applied Econ., JIdy, 1993
The interest rate coefficient for Saratoga
sales was negative, but not significantly different
from zero at the 10 percent level. A one percent
increase in the interest rate will on]y decrease
The dummy
yearling prices by .007 percent.
variable for the Tax Reform Act of 1986 has a
parameter estimate of -0.17 which is significant at
the 10 percent level, implying that Saratoga yearling
prices fell 17 percent due to the tax change -almost ten percentage points lower than the effect
on Keeneland prices.
Apportioning Changes to Exogenous Variables
Table 2 indicates the change in yearling
prices for the highest-priced quartile at Keeneland
between 1980 and 1984, and 1984 and 1990 into
changes due to the four exogenous variables -- the
purse rate, exchange rate, interest rate, and change
in the tax laws4. Between 1980 and 1984, two
relevant exogenous variables moved to increase
yearling prices, while one (the interest rate) moved
slight]y to decrease yearling prices. The two effects
dominated the one, with the stronger positive impact
coming from the exchange rate depreciation. The
total increase in median yearling prices between
those early years was $412,000 (or 154 percent),
with the purse rate accounting for $168,000 (or 41
percent of the change) and the exchange rate
accounting for $248,000 (or 60 percent of the
change). ‘The sum of the exchange rate and purse
rate effect was greater than 100 percent because the
interest rate moved slightly to reduce thoroughbred
prices.
Table 2 shows that the tax law change in
1986 was a major negative factor influencing
yearling prices, though the exchange rate also had
a large negative impact. The 1986 tax change alone
caused yearling prices to fail $185,000 (or 27
percent). The exchange rate appreciation caused
prices to fall by $135,000 (or 20 percent), while
interest rate changes helped yearling prices fall by
$27,000 (or 4 percent).
Conclusion
The estimates for Keeneland and Saratoga
provide evidence that macroeconomic factors,
exchange rate and tax law changes, have
significantly influenced the thoroughbred industry.
A major factor which helped yearling prices
increase in the early 1980s (the exchange rate)
turned against the yearling market in the late 1980s.
Another serious price-reducing event was the tax
law change.
Phenotypic factors notwithstanding, the
effects of these macroeconomic variables imply that
the thoroughbred industry is influenced by monetary
and fiscal policies. A weak dollar relative to
currencies of leading buyers, lead to higher yearling
prices. Higher levels of interest rate (for the
Saratoga market) and the Tax Reform Act of 1986
have also had a negative effect on the industry. The
Tax Reform Act of 1986, in particular, had a drastic
negative effect on the industry. The tax law could
be changed through future actions and there are
moves to increase purse rates through state
programs, but it appears that the thoroughbred
industry will probably have to rely on a major
dollar depreciation to turn yearling prices around.
The effect of the exchange rate on the
thoroughbred industry enhances our understanding
of its effect on a specific commodity or product.
The yearling market presents a unique opportunity
to study the impacts of an exchange rate change
because of the obvious presence of foreign buyers
at the auction. If a product relies heavily on foreign
markets, the exchange rate can play a major role in
product demand and price.
172
Karungu, Reed and Tvedt:
Macroeconomic
Factors
Table 2. Apportioned Change in Yearling Prices, Highest Quartile, 1980-84 and 1984-90, in Thousands
of Dol Iars
Year
Total
Effecta
Purse
Rate
Exchange
Rate
Interest
Rate
Tax
Change
1980
to
1984
$412
$168
$248
-$4
.-
1984
to
1990
-$209
$137
-$135
-$27
-$185
‘The total effect measures the total predicted change in yearling prices by the model for the highest quartile
horses. The purse rate, exchange rate, interest rate and tax change effects measure the predicted change in
yearling prices from changes in those variables, respectively,
References
Bessler, David A. “Relative Prices and Money: A Vector Autoregression on Brazilian Data.” Amer. J. Agr.
Econ. 66(1984):25-30.
Chambers, Robert G. “ Agricultural and Financial Market Interdependence in the Short-Run.” Amer. J. Agr.
Econ, 6(1984): 12-24.
Chambers, Robert G., and Richard E. Just. “ Effects of Exchange Rate Changes on U.S. Agriculture: A
Dynamic Analysis. “ Amer. J. Agr. Econ. 63(1980): 32-46.
Devados, S., and William H, Meyers, “Relative Prices and Money: Further Results of the United States.”
Amer. J. Agr. Econ. 6(1984): 838-42.
Johnson, Paul R., Thomas Grennes, and Marie Thursby “ Devaluation Foreign Trade Controls and Domestic
Wheat Prices.” Amer. J. Agr. Econ. 59(1977): 619-27.
Kost, William “ Effects of Exchange Rate on Agricultural Trade.” Agr. Econ. Res. 28(1976): 99-106,
Lawrence, Robert G. “All About Purses, “ The Thoroughbred Record, Aug. 1989, pp.886-889
Longmire, J., and Morey, A, “Strong Dollar Dampens Demand for U.S. Farm Exports.” Foreign
Agricultural Economics Report No. 193, Economic Research Service, United States Department of
Agriculture, Washington D.C. (1983),
J. Aw. and Applied Ecan., July, 1993
1’73
Newberry, James H. Jr., Legal Aspects of Horse Farm Operations, University of Kentucky College of Law,
Continuing Legal Education Monograph Series. 1990.
Orden, David, “ Agriculture, Trade, and Macroeconomics: The U.S. Case.” Journal of Policy Modelling
8 ( 1986): 27-51.
Schuh, Edward G., “ The Exchange Rate and U.S. Agriculture.” Amer. J. Agr. Econ. 56 (1974): 1-13.
The Blood-Horse: A Weekly Publication of the Thoroughbred Owners and Breeders Association, Lexington,
KY.: Blood-Horse, Inc., various issues.
Thoroughbred Times, Lexington, Ky: Thoroughbred Publications, January 1990.
Footnotes
1 All thoroughbreds become yearlings on the first January 1 after they are foaled. Each thoroughbred has
January 1 as its official birthday.
2 Using quartiles increases the degrees of freedom (number of observations) available for analysis. An
anal ysis of the top quartile alone doesn ‘t have enough error degrees of freedom. The median price within
a quartile will reduce the impact of phenot ypic factors because variation among the high-priced horses
associated with very specific sires will be eliminated, Some argue that yearling prices have fallen recently
because there are no longer yearlings sired by Northern Dancer and Nijinsky II. This type of effect is
reduced by using the median price.
~ The exchange rate index for each year is calculated using the following formula:
$~wi
i=l fJ
where i is a country subscript, EOis the exchange rate in 1982 (the base year) and Wj is the weight of that
country’s exchange rate for the index. An indexing form is needed because the units of Ei vary by country.
The weights for each year are determined by the percentage of yearlings purchased by the five leading
thoroughbred importing countries.
4 1984 is chosen for comparison because it is the year that thoroughbred prices peaked. Price changes for
the other quartiles would be smaller absolute changes, but the relative effects of macroeconomic variables
would be identical.